4491 speakers participated in the recording and conducted face-to-face communication in a natural way. No topics are specified, with a wide range of fields; the voice was natural and fluent, in line with the actual dialogue scene. Text is transferred manually, with high accuracy. Format：16kHz, 16bit, uncompressed wav, mono channel Environments：quiet indoor environment, without echo Recording content：no topic is specified, and the speakers make dialogue while the recording is performed Demographics：4,491 speakers, 63% of which are female. Annotations：annotating for the transcription text, speaker identification and gender Device：Android mobile phone, iPhone Language：Mandarin Applications：speech recognition;  voiceprint recognition. Accuracy rate：97%

Creator(s)

-

Distributor(s)

ELRA

Right Holder(s)

-